We use RedHat JBPM's KIE server as our rules engine. Our application gets around 1000 TPS in pick hours. We also have built multiple KIE container which holds rules for different domains. Since KIe-server allows to update those rules on the fly, we use kie-server's scanner API to the same.
We are facing this issue when KIE container update is in progress or few miliseconds/seconds after updates, CPU spikes up and transaction time for real time requests would start increasing and APP has to timeout on those requests. This happens very frequently and pattern is always during or soon after container updates. Few calls fails due to this issue and looking for help to find out root cause here. We use KIE version 7.73.0.Final.
Any suggestion to find out root cause on what blocks real time requests during container updates or any best practices for KIE container updates that we might not be following already would be helpful. Thanks.
We tried to enable Trace level logging and see what could be the root cause but no luck. We also tried upgrading version two times already but no luck there.